surebackup proxy starting with no routes

[iltmpz]

Hi all!

So, i have an exchange 2010 farm contains of 6 servers: 2 cas, 2 storage in dag, 2 more storage in dag. And 4 domain controllers.
I use veeam B&R 7.0

I have 2 subnets: 192.168.30.0/24 as client subnet and 192.168.31.0/25 as server one.
Both domain controlles and mail storages located in server subnet only.
CAS-server has 2 interfaces: in server and client subnets.

These subnets use different vlans.

I made replicas for 2 domain controllers, 2 cas and 2 storage to the single server.
Now I try to configure surebackup for that.

I made application group named ex_test, virtual lab called ex_test and surebackup task called ex_test.

My config for virtual lab:
name: ex_test,
host&datastore: i use one for all,
proxy: checked "use..."
Production network: i have to select server subnet and any free static ip in my server subnet, right? I tried eighter static configuration or DHCP. For example: 192.168.31.107
networking: advanced single host (while i use the one physical esx for the whole backup exchange farm)
isolated networks: added both "ClientNet - ex_test ClientNet" and "ServerNet - ex_test ServerNet", including proper vlans for each one.
network settings: I set appliance ip to 192.168.30.1 and 192.168.31.1 for client and server networks (same as gateways in my real networks).
what to select as masquerade ip? Looks like doesn't matter, i leaved default 192.168.254.x and 192.168.253.x(/25). These subnets are never used in my infrastructure.

static mapping: just leaved empty.
Finish.

Not when i start my surebackup task it runs proxy vm called ex_test, succeeded and try to run dc1 (domain controller).
And here is my first problem:
proxy and dc1 are both attached to the ex_test ServerNet, but can't ping each other.
When i open the proxy unix console (by root/ex_test_r) i can ping the whole my public network, but could not see running dc1 inside "ex_test ServerNet" network.
When i type "route" in console, i see only 1 default route via eth0 interface (my public network).

Then both proxy and dc1 vms are running, but veeam task shows dc1 status as starting. Heartbeat: success. Ping: in progress.
And ping is really not working.

When i type manually in proxy console: route add -net 192.168.31.0/25 eth2, then dc1 and proxy both can ping each other.
But surebackup task still hangs in state "Ping - in progress".

I find this: http://www.veeam.com/blog/automate-reco ... ackup.html

When you start a virtual lab, a static route to a masquerade network is added automatically on the Veeam Backup & Replication server.

I didn't see anything about my masquerade networks. Only 1 default route to production network.

So, where can i add the static routes to let proxy and starting replicas see each others? Or what esle need i do to let both heartbeat and ping tests be succeed?

[Andreas Neufert]

OK hard to find the main reason for this... But let me guess 2 things...

1) Is public network interface from vlab appliance and B&R Management on same subnet/network? If not... do so.
2) At vlab configuration wizard. Did you changed the mapped network name of second added network?
e.g. you have network "LAN" and "DMZ" in production. Wizard automatically add "LAN" - "VLAB LAN" mapping. When you add second mapping the network name needs to be edited by clicking direct on the line and change "DMZ" - "VLAB LAN" to "DMZ" - "VLAB DMZ".

[iltmpz]

1) Is public network interface from vlab appliance and B&R Management on same subnet/network? If not... do so.

Well, vmware management network called "Management Network" and is not allowed to be added as production network for virtual lab.
I used network called ServerNet - the same subnet as "Management Network", mean the same ip range: 192.168.31.1/25.

2) At vlab configuration wizard. Did you changed the mapped network name of second added network?
e.g. you have network "LAN" and "DMZ" in production. Wizard automatically add "LAN" - "VLAB LAN" mapping. When you add second mapping the network name needs to be edited by clicking direct on the line and change "DMZ" - "VLAB LAN" to "DMZ" - "VLAB DMZ".

Well, first network: production name: "ClientNet", isolated: "ex_test ClientNet"
second one: production name: "ServerNet", isolated: "ex_test ServerNet"

I just tried to make new vlab, new appgrp, new task, called ex_test1 with only dc1 included. I have the same result:
just no route between proxy and dc1 vms there.
So, heartbeat test succeeds, ping test fails.

And now i manually added route to proxy for dc1 behind my proxy: "route add -host 192.168.31.3 eth2" before my dc1 replica have been started.
Then both heatbeat and ping tests were succeed. So, the only question now, is why didn't it add the required route automatically? Where can i add this host using vlab wizard or w/e?

[Andreas Neufert]

Please send me an email to andreas.neufert at veeam.com
I will organize an webex with you and we can check the configuration and environment together.

[iltmpz]

Thank you all, the problem has been solved.

I recreated the task, increased execution time and it works successfully.
Now my total execution time for 6 virtual machines in surebackup task is about 2 hours. But it works.

[Andreas Neufert]

Do you use any Dedup Storage as Backup Targets (StoreOnce/Datadomain/DXI/Windows Dedup)?